Update app.py

app.py

@@ -38,7 +38,7 @@ device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 peft_model_id = "sooolee/flan-t5-base-cnn-samsum-lora"
 config = PeftConfig.from_pretrained(peft_model_id)
 tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)
-model = AutoModelForSeq2SeqLM.from_pretrained(config.base_model_name_or_path, device_map='auto') # load_in_8bit=True,
+model = AutoModelForSeq2SeqLM.from_pretrained(config.base_model_name_or_path, device_map='disk_offload') # load_in_8bit=True,
 model = PeftModel.from_pretrained(model, peft_model_id, device_map='auto')
 
 def summarize(video_id):